The EMBO Journal Vol. 21 No. 24 pp. 6721±6732, 2002 Physiological concentrations regulate binding and catalysis of adenylyl cyclase exotoxins

Yuequan Shen1, Young-Sam Lee2, the central helix and the calcium-induced changes in Sandriyana Soelaiman1, Pamela Bergson1,3, surface properties enable CaM to bind and modulate a Dan Lu1, Alice Chen1, Kathy Beckingham4, diverse array of physiologically important , such Zenon Grabarek5, Milan Mrksich2 and as adenylyl cyclase, phosphodiesterase, Wei-Jen Tang1,3,6 synthase, kinase and phosphatase, receptor and ion channel (Eldik and Watterson, 1998; DeMaria et al., 1Ben-May Institute for Cancer Research, 2Department of Chemistry, 2001). Consequently, CaM is involved in many intra- 3 and Committee on Neurobiology, The University of Chicago, cellular processes including control of transcription, ion Chicago, IL 60637, 4Department of Biochemistry and Cell Biology, Rice University, Houston, TX 77251 and 5Boston Biomedical ¯uxes, signal transduction, vesicular transport and cyto- Research Institute, Watertown, MA 02472, USA skeleton functions (Deisseroth et al., 1998; Eldik and Watterson, 1998). 6Corresponding author e-mail: [email protected] Structural and biochemical analyses have provided insights into CaM-dependent regulation of some target Edema factor (EF) and CyaA are calmodulin (CaM)- enzymes (Hoe¯ich and Ikura, 2002; Meador and Quiocho, activated adenylyl cyclase exotoxins involved in the 2002). The best-known activation mechanism is the pathogenesis of anthrax and , respect- release of autoinhibition exempli®ed by myosin light ively. Using spectroscopic, enzyme kinetic and surface chain kinase (MLCK) and CaM kinase II (CaMKII). In this plasmon resonance spectroscopy analyses, we show model, CaM is not associated with the target enzyme at the that low Ca2+ concentrations increase the af®nity of resting Ca2+ levels, and the enzyme's catalytic activity is CaM for EF and CyaA causing their activation, but blocked by its own autoinhibitory domain (AID). Upon higher Ca2+ concentrations directly inhibit catalysis. increase in the intracellular Ca2+ concentration, the Both events occur in a physiologically relevant range Ca2+±CaM complex binds to an amphipathic a-helix that of Ca2+ concentrations. Despite the similarity in Ca2+ partially overlaps the AID. This presumably causes a sensitivity, EF and CyaA have substantial differences conformational change that disrupts the interaction of the in CaM binding and activation. CyaA has 100-fold AID with the catalytic domain, resulting in kinase higher af®nity for CaM than EF. CaM has N- and activation. A recent structure of the complex of CaM C-terminal globular domains, each binding two Ca2+ with the intracellular domain of the small conductance 2+ ions. CyaA can be fully activated by CaM mutants Ca -activated potassium channel, Ik(Ca), reveals a new with one defective C-terminal Ca2+-binding site or by activation mechanism (Schumacher et al., 2001). In this either terminal domain of CaM while EF cannot. EF case, CaM is constitutively bound to Ik(Ca). Calcium consists of a catalytic core and a helical domain, loading allows CaM to act as a clamp to induce Ik(Ca) and both are required for CaM activation of EF. dimerization, which leads to an allosteric change and Mutations that decrease the interaction of the helical increased ion conductivity. domain with the catalytic core create an enzyme with Several pathogenic , such as those that cause higher sensitivity to Ca2+±CaM activation. However, anthrax (Bacillus anthracis), whooping cough ( CyaA is fully activated by CaM without the domain pertussis) and cholera (), secrete corresponding to the helical domain of EF. that increase the cAMP concentration in the host cells to a Keywords: adenylyl cyclase exotoxin/anthrax edema pathological level (Drum et al., 2002). Edema factor (EF) factor/Ca2+±calmodulin/CyaA/enzyme activation and CyaA are adenylyl cyclase toxins produced by B.anthracis and B.pertussis, respectively (Ladant and Ullmann, 1999; Mock and Fouet, 2001). Both EF and CyaA are activated in the host cell by CaM. EF is Introduction responsible for the massive edema seen in cutaneous Calcium serves as a diffusible second messenger in anthrax and impairs the function of and response to extra- and intracellular signals, and calmodulin in systemic (Hoover et al., 1994). (CaM) is a key calcium sensor (Eldik and Watterson, CyaA is important for the bacterial colonization of the 1998). CaM has two globular domains, each consisting of respiratory tract, in part owing to its ability to induce two helix±loop±helix calcium-binding motifs (Babu et al., of (Khelef et al., 1993; Weingart 1985). The domains are linked by a ¯exible a-helix that is and Weiss, 2000). partially unfolded in solution (Barbato et al., 1992). The CaM-activated adenylyl cyclase domain resides in Calcium induces a transition in both domains from a the C-terminal 510 region of EF. The closed conformation with highly negatively charged molecular structures of this domain with and without surface to an open conformation with a large, exposed CaM were solved recently. These structures have provided hydrophobic pocket (Finn et al., 1995). The ¯exibility of a new model of CaM binding and activation of a target

ã European Molecular Biology Organization 6721 Y.Shen et al.

Fig. 1. The effect of calcium ions and CaM on the adenylyl cyclase activities of EF and CyaA-N. Adenylyl cyclase assays were performed in the presence of 1 nM EF (A) and 0.7 nM CyaA-N (B) under 10 mM CaM (®lled circles), 0.1 mM CaM (open circles) and 1 nM CaM (®lled triangles, CyaA-N only) at increasing [Ca2+]. They were also performed at 0.1 mMCa2+ (®lled circles), 0.3 mMCa2+ (open circles) and 1.0 mMCa2+ (®lled triangles) at increasing [CaM] in the presence of 1 nM EF (C), and 0.7 nM CyaA-N (D). Maximal adenylyl cyclase activities (100%) for EF in the calcium titration are 1140 s±1 (10 mM CaM) and 228 s±1 (0.1 mM CaM) (A) and those for CyaA-N are 1465 s±1 (10 mM CaM), 713 s±1 (0.1 mM CaM) and 556 s±1 (1 nM CaM) (B). Means 6 SE are representative of at least two experiments.

enzyme (Drum et al., 2002; Hoe¯ich and Ikura, 2002; In this paper, we show that the physiological intra- Meador and Quiocho, 2002). Unlike the CaM complexes cellular calcium concentration not only dramatically with MLCK and CaMKII, in which CaM takes a compact increases the CaM af®nity of both EF and CyaA, form, in the complex with EF, CaM has an extended enhancing the activation of adenylyl cyclase activity, but conformation and makes extensive contacts (~6000 AÊ 2) also directly interferes with binding of the catalytic metal, with four discrete regions of EF. This interaction induces a inhibiting catalysis. We also show that the interaction of 15 AÊ translation and 30° rotation of a 15 kDa helical the helical domain of EF with CaM plays a key role in EF domain of EF resulting in stabilization of a 12 amino acid- activation. long loop. This loop contains several amino acid residues crucial for catalysis, and it is disordered in the structure of EF alone so that the catalytic site is open and incomplete. Results CaM-induced conformational changes stabilize this loop Physiological calcium and CaM concentrations are to enclose and complete the catalytic site, achieving over required for the optimal activation of EF 1000-fold enhancement in the catalytic rate of EF. The activation of EF by CaM can be greatly reduced by the There is considerable cross talk between the signaling addition of EGTA, a calcium chelator (Leppla, 1984). pathways of calcium and cAMP, two key intracellular However, it is not clear whether physiological concentra- messengers. An increase in intracellular cAMP levels can tions of calcium could modulate the activation of EF by elevate intracellular Ca2+ by the activation of calcium- CaM. While the estimated total CaM concentration inside permeable channels, such as L-type calcium channels, cells is ~1±10 mM, free CaM concentration is signi®cantly nicotinic acetylcholine receptors and cyclic nucleotide lower since it is associated with apo-CaM-binding proteins gated channels (Cooper et al., 1995). On the other hand, such as GAP-43 and RC3 (Gerendasy, 1999; Jurado et al., Ca2+±CaM can decrease the cAMP level by activation of 1999). We therefore measured adenylyl cyclase activity of phosphodiesterase, or increase it by activating speci®c EF at a broad range of calcium concentrations (1 nM to isoforms of adenylyl cyclase (type I, III and VIII; Eldik 10 mM) in the presence of either 0.1 or 10 mM CaM, and Watterson, 1998; Hanoune and Defer, 2001). Little is representing the low and high ends of the intracellular free known regarding the effects of physiological Ca2+ con- CaM concentration range (Figure 1A). We found that, centrations on the adenylyl cyclase activity of EF and at both CaM concentrations, adenylyl cyclase activity CyaA. This question is especially intriguing in view of the exhibited a bell-shaped curve with optimal activity at the fact that Ca2+ ions were found only in the C-terminal physiologically relevant calcium concentrations (0.1± domain of CaM in the structure of EF±CaM complex. 0.5 mM). While <5-fold change in the adenylyl cyclase

6722 Activation of EF and CyaA by Ca2+±calmodulin activity of EF at the 0.03±10 mM calcium concentrations anthraniloyl group in 2¢d3¢ANT-ATP is covered by a was observed in the presence of 10 mM CaM, the catalytic loop called switch B (Figure 2C and D). Switch B enzymatic activity of EF was sharply altered by the is not visible in the EF alone structure, but it is stabilized same calcium concentrations when 100-fold less CaM by switch C and becomes ordered upon CaM binding. The (0.1 mM) was used. This requirement of calcium and CaM anthraniloyl group in 2¢d3¢ANT-ATP is aligned with the for the optimal activation of EF was observed with both phenyl group of F586, a residue in switch B. Consistent the physiological Mg2+ concentration (0.5 mM; Supple- with the structural model, the mutation of F586 to alanine mentary ®gure 1, available at The EMBO Journal Online) greatly reduced the ability of EF to enhance the ¯uores- and the Mg2+ concentration optimal for EF activity cence of 2¢d3¢ANT-ATP (Figure 2A). The F586A muta- (10 mM; Figure 1A). This data suggests that the optimal tion had a relatively small effect on EC50 and Vmax values activation of EF requires the physiological calcium for CaM activation and Km for substrate binding (the ±1 concentrations, particularly when CaM is limited. EC50,CaM, Vmax and Km,ATP values are 2 nM, 718 s and The 1706 amino acid CyaA is a bi-functional protein 0.7 mM for wild-type EF and 2.5 nM, 370 s±1 and 1.0 mM with both adenylyl cyclase and hemolytic activities. The for EF-F586A). N-terminal 45 kDa domain of CyaA has CaM-activated Based on ¯uorescence data on calcium-dependent CaM- adenylyl cyclase activity and its C-terminal 140 kDa binding to EF, we predicted that physiological calcium domain contains a transmembrane region, 42 -rich concentrations can alter the concentration of CaM required motifs that bind calcium and a hemolytic domain. We to achieve half maximal activation of EF (EC50). To test expressed and puri®ed the functional N-terminal adenylyl this prediction, we examined the ability of CaM to activate cyclase domain (aa 1±393, referred to here as CyaA-N) of EF at 0.1, 0.3 and 1 mM calcium (Figure 1C). Our results CyaA without denaturant. We then addressed whether the showed that with increasing calcium concentrations, the activation of the adenylyl cyclase domain of CyaA also EC50 values for CaM activation of EF decreased from depends upon calcium and CaM concentrations. In con- 2 mM to 50 nM. Calcium concentrations >1 mM did not trast to EF, the activity of CyaA-N was optimally cause further signi®cant decrease of the EC50 values stimulated by CaM at 1 nM to 0.1 mM calcium with either (Supplementary ®gure 3). These results agree with the 0.1 or 10 mM CaM, whereas its activity was signi®cantly notion that physiological calcium concentrations can inhibited at calcium concentrations >0.1 mM (Figure 1B). affect the af®nity of EF to CaM and highlights the free With only 1 nM CaM, the adenylyl cyclase activity of energy coupling between Ca2+ and EF binding to CaM. CyaA-N exhibited a bell-shaped curve in the calcium We then studied the interaction of EF with immobilized concentrations range from 1 nM to 10 mM. This data CaM by surface plasmon resonance (SPR) spectroscopy suggests that, in contrast to EF, CyaA-N is optimally (Figure 3). We immobilized CaM to a self-assembled activated at resting calcium concentrations with free CaM monolayer using a recently reported active site-directed ranging from 0.1 to 10 mM, and its activity is reduced with immobilization method (Hodneland et al., 2002). CaM elevated intracellular calcium concentrations. A similar was fused to the C-terminal end of cutinase from the observation was found using the full-size CyaA (Gentile fungus Fusarium salani and the cutinase±CaM fusion et al., 1990). protein (cut-CaM) was covalently bound to a gold-coated glass surface previously coated with 4-nitrophenyl- Physiological calcium concentrations modulate the phosphonate ligand, a transition state substrate analog af®nity of CaM for EF and CyaA-N of cutinase. Cut-CaM activated EF with 5-fold reduced Calcium concentration may affect the interaction between af®nity (Supplementary ®gure 4). The SPR analysis CaM and EF. To address this question, we took advantage showed that the fraction of immobilized CaM bound of the observation that the emission intensity of the to EF increases at elevated calcium concentrations ¯uorescent ATP analog 2¢-deoxy,3¢-anthraniloyl ATP (Figure 3A), providing further evidence that calcium (2¢d3¢ANT-ATP) increases ~3-fold upon binding to the ions enhance the binding of these two proteins. These CaM±EF complex (Figure 2A; Sarfati et al., 1990). From experiments reveal that CaM binds EF with a Kd of 20 nM the equilibrium titration monitored by ¯uorescence, we at 10 mM calcium, in agreement with the reported af®nity found that the Kd for the binding of 2¢d3¢ANT-ATP to (Figure 3E; Drum et al., 2002). The SPR analysis also CaM±EF is 1 mM, and the binding requires physiological showed that calcium ions predominantly increased the rate calcium concentrations (0.05±2 mM; Figure 2A and B). In of association (ka1) and that of the conformational view of the observation that 2¢d3¢ANT-ATP is a potent transition (K2) rather than decreasing the dissociation rate. inhibitor of CaM±EF (Sarfati et al., 1990), we then asked To ensure that the SPR experiments re¯ected the whether this compound binds to the enzyme in a manner binding of CaM to EF, we examined EF-K525A, an EF similar to the normal substrate ATP. To answer this mutant with a point mutation changing 525 to question and to explain the mechanism of the calcium/ alanine. The structure of the EF±CaM complex reveals CaM-dependent enhancement of 2¢d3¢ANT-ATP ¯uores- that lysine 525 of EF forms a salt bridge with cence, we solved the structure of the EF±CaM± glutamate 114 of CaM (Drum et al., 2002). Our SPR 2¢d3¢ANT-ATP complex (Table I; Figure 2C and D). analysis showed that at the highest calcium concentration There is a clear representation of 2¢d3¢ANT-ATP in the assayed, EF-K525A achieved a signal comparable with simulated annealing omit map of the EF±CaM±2¢d3¢ANT- wild-type EF at low levels of calcium (Figure 3B). ATP structure (Supplementary ®gure 2). The portions of Consistent with our biochemical analysis, the af®nity of the structure comprised of EF±CaM and the 3¢d-ATP EF-K525A for CaM was two orders of magnitude lower moiety of 2¢d3¢ANT-ATP are virtually identical to the than that of the wild-type EF at 100 mM calcium structure of EF±CaM±3¢d-ATP (Drum et al., 2002). The (Figure 3E; Drum et al., 2002).

6723 Y.Shen et al.

Fig. 2. The binding of 2¢d3¢ANT-ATP to EF. (A) Equilibrium titration of 2¢d3¢ANT-ATP±CaM with EF and EF-F586A. (B) Calcium titration of ¯uorescence enhancement by EF±CaM. 2¢d3¢ANT-ATP was added to a ®nal concentration of 0.5 mM and the indicated free calcium concentrations were achieved by buffering with 10 mM EGTA. lexc = 320 nm and the optimal ¯uorescence emission of EF±CaM±2¢d3¢ANT-ATP (412 nm) was normalized to give the fold of enhancement. (C) Secondary structure of EF±CaM±2¢d3¢ANT-ATP in comparison with EF alone. (D) The active site of EF in the presence and absence of CaM and 2¢d3¢ANT-ATP.

We also examined whether calcium concentrations can dissociation rate from cut-CaM (Kd1) was signi®cantly alter the EC50 value for the activation of CyaA-N by CaM reduced for CyaA-N as compared with EF (Figure 3E). (Figure 1D). We found that calcium concentrations from 0.1 to 1 mM reduced the EC50 value for CaM activation of Calcium ions directly inhibit adenylyl cyclase CyaA-N from 100 to 2 nM, and calcium concentrations activity of EF and CyaA-N beyond 1 mM did not further reduce EC50 values (Supple- In the EF±CaM structure, only the calcium-binding sites in mentary ®gure 3). To evaluate the binding of CaM to the C-terminal domain of CaM are loaded with calcium, CyaA-N directly, we then examined the interaction of whereas the N-terminal domain does not contain Ca2+ and CyaA-N with cut-CaM using SPR (Figure 3D). Cut-CaM adopts the conformation of apo-CaM (Drum et al., 2002). activated CyaA-N with an EC50 similar to that of wild-type As described above, we found that calcium concentrations CaM, and with 2-fold higher Vmax (Supplementary >1 mM signi®cantly reduce the adenylyl cyclase activities ®gure 4). The SPR analysis showed that CyaA-N reached of both EF and CyaA-N. We hypothesized that the levels of binding to cut-CaM comparable with those inhibitory effect of calcium might be caused by calcium observed for EF, but at much lower calcium concentra- binding to the N-terminal domain of CaM and a subse- tions (Figure 3). Although the rates of association (Ka1) quent change in the interaction between this domain and and of the conformational transition (K2) for the cut-CaM EF. To test this hypothesis, we examined the ability of a binding were similar between EF and CyaA-N, the mutated form of CaM, CaM41/75, to activate EF and

6724 Activation of EF and CyaA by Ca2+±calmodulin

CyaA-N (Figure 4A and C). CaM41/75 has cysteines Table I. Statistics of the EF±CaM±2¢d3¢ANT-ATP complex data set replacing residues 41 and 75, capable of forming a Data collection disul®de bond and locking the N-terminal domain of CaM in the closed conformation, irrespective of calcium (Tan Beamline APS, 14-BM-C Space group I222 et al., 1996). If calcium binding in the N-terminal domain Unit cell (AÊ ) of CaM and the resulting conformational change cause the a 116.92 reduced activation of EF and CyaA-N by CaM, then the b 167.92 activation by CaM41/75 should not decrease at high Ca2+ c 341.74 Resolution (AÊ ) 3.6 concentrations. Our result showed that CaM41/75 acti- Completeness (%) 99.4 vates EF and CyaA-N in a manner indistinguishable from Redundancya 6.85 the wild-type CaM (Figure 4A and C; data not shown for R (%)b 8.3 sym Ca2+ titration). Thus, the inhibition for the activities of EF I/s 11.3 and CyaA-N by Ca2+ is unlikely to be mediated through Re®nement the Ca2+ binding to the N-terminal domain of CaM. We then examined whether calcium bound to the R (%)c R (%)d R.m.s. (AÊ ) R.m.s. (°) cryst free bond angle substrate in place of Mg2+ could cause inhibition of the 28.1 30.7 0.012 1.8 catalytic activity. We found <0.2% of adenylyl cyclase a 2+ Nobs/Nunique. activity for both EF and CyaA-N when Ca ±ATP was b Rsym = Sj|±Ij|/S, where Ij is the intensity of the jth re¯ection used as the substrate as compared with the magnesium± and is the average intensity. c ATP (data not shown; Labruyure et al., 1990). Thus, Rcryst = Shkl|Fobs ± Fcalc|/ShklFobs. 2+ d Ca ±ATP is a poor substrate for both EF and CyaA. To Rfree, calculated the same as for Rcryst but on the 5% data excluded from the re®nement calculation. understand the molecular mechanism behind calcium- mediated inhibition, we analyzed the dependence of

Fig. 3. The effect of calcium ions on the interaction of EF and CyaA-N with immobilized CaM by SPR sensorgram analysis. Sensorgrams were recorded in the presence of 1.4 mM wild-type EF (A), 1.3 mM EF-K525A (B), 0.37 mM CyaA-N (D) and 8.0 mM EF-triple mutant (C) at the indicated free calcium concentrations. (E) Kinetic analysis of SPR sensorgrams based on a `two-state conformational change' model.

ka1 K2 à E ‡ CaM „ E Á CaM „ E Á CaM K1 ˆ kd1=ka1 and Kd;app ˆ K1= 1 ‡ K2† kd1 At the lowest calcium concentration, signal from the bound EF could be reduced to the base line with the wash using the same calcium concentration. However, a fraction of the signal could not be reduced at the higher calcium concentrations until the buffer with the lowest calcium concentration was used in the wash step. The signal from the bound CyaA-N could not be reduced to the base line without a prolonged wash.

6725 Y.Shen et al.

Fig. 4. The activation of EF and CyaA-N by wild-type CaM and two series of CaM mutants. EF (1 nM) and CyaA-N (0.7 nM) were used for an adenylyl cyclase activity assay in the presence of 0.1 mM free Ca2+ with CaM mutants CaM 41/75 and 85/112. Each mutant has two cysteine mutations to lock either the N- or the C-terminal domain of CaM in the closed conformation (A and B). The same concentrations of EF and CyaA-N were used for an adenylyl cyclase activity assay in the presence of 1 mM free Ca2+ with CaM mutants B1Q, B2Q, B3Q and B4Q. Each mutant has a mutation inactivating one of four calcium-binding sites (C and D). Means 6 SE are representative of at least two experiments. adenylyl cyclase activity of EF and CyaA-N on magne- 10-fold compared with that of the wild-type EF sium at 0.1, 0.3 and 1 mM calcium (Figure 5A and B). We (Figure 5C). Similar reduction in sensitivity to calcium observed a bell-shape curve for the magnesium require- inhibition was also observed with Ca2+ concentrations ment with the optimum between 1 and 20 mM for both EF ranging from 0.1 to 1 mM (Figure 5D). Together our data and CyaA-N. In the magnesium concentrations ranging suggest that a magnesium ion binds to the catalytic metal from 0.1 to 10 mM, we found that the higher the calcium site of EF and that calcium interferes with the binding of concentration, the greater the magnesium concentration this magnesium ion. required for maximal activity. Assuming that the calcium titration curves in Figure 1A and B were composed of a calcium-dependent activation, ®t by a simple rectangular EF and CyaA-N have different requirements for hyperbola, and a calcium-dependent inhibition using their activation by CaM Our SPR and kinetic analyses demonstrate that CaM binds 2+ 2+ CyaA-N with 20- to 100-fold higher af®nity than it does v = Vmax 3 ([Ca ]/(EC50 + [Ca ])) 3 2+ EF. In order to identify the roles of the different structural (1/(1 + ([Ca ]/IC50))) elements of CaM in activating these two enzymes, we we estimated the half maximal concentration for calcium examined their activation by three sets of CaM mutants. inhibition (IC50) of EF and CyaA-N to be ~0.3±1 mM, CaM 85/112 has cysteine substituted for residues 85 and suggesting that the calcium inhibition of EF and CyaA is 112, resulting in the C-terminal domain of CaM being mediated by the high nanomolar af®nity calcium-binding locked in the closed conformation by a disul®de bond (Tan site(s). et al., 1996). We found that CaM85/112 activated EF In our EF±CaM±3¢dATP structure, the catalytic site of poorly, while it could activate CyaA-N with an increased EF has one catalytic metal ion, which is coordinated by EC50 value and 50% decrease in Vmax (Figure 4A and C) two aspartates, D491 and D493 (Drum et al., 2002). We also used a set of CaM mutants (Q-series), each of Histidine 577 of EF forms hydrogen bonds with these which had a mutation in one of four calcium-binding sites aspartates in addition to contacting the catalytic metal (Maune et al., 1992), to probe the contribution of (Figure 2D). We found that EF-H577N, an EF mutant with individual calcium-binding sites to the activation of EF asparagine substituted for histidine 577, had adenylyl and CyaA-N (Figure 4B and D). In this series of CaM cyclase activity reduced by two orders of magnitude even mutants, the conserved glutamate at position 12 of each though this mutant had normal ATP-binding activity and calcium-binding site was mutated to glutamine, a change CaM activation (Drum et al., 2002). We used the metal that effectively eliminates calcium binding. As observed sensitivity of EF-H577N to determine whether H577 plays previously, the mutation at either site 3 or 4 severely a role in the binding of calcium and magnesium ions. We reduced the Vmax and EC50 values for CaM activation of found that, with Mg2+ ranging from 0.1 to 10 mM, the EF, whereas those at either site 1 or 2 had a minimal effect sensitivity of EF-H577N to magnesium was reduced 5- to (Figure 4B; Drum et al., 2000). In contrast, only the

6726 Activation of EF and CyaA by Ca2+±calmodulin

Fig. 5. Effect of calcium and magnesium ions on adenylyl cyclase activity of EF and CyaA-N (A and B) and an EF mutant, EF-H577N (C and D). Adenylyl cyclase activity assays were performed with 10 mM CaM at 0.1 mMCa2+ (®lled circles), 0.3 mMCa2+ (open circles) and 1.0 mMCa2+ (®lled triangles) in the presence of 1 nM EF (A), and 0.7 nM CyaA-N (B). To analyze the mutant form of EF, adenylyl cyclase activities were measured with 10 mM CaM and 1 nM EF or 66 nM EF-H577N with either 0.3 mMCa2+ (C) or 10 mM Mg2+ (D). Both were buffered by 10 mM EGTA. Maximal activities for EF in the magnesium titration are 1616 s±1 (0.1 mMCa2+), 1074 s±1 (0.3 mMCa2+), and 1009 s±1 (1.0 mMCa2+) (A) and those for CyaA-N are 2106 s±1 (0.1 mMCa2+), 1674 s±1 (0.3 mMCa2+) and 1385 s±1 (1.0 mM) (B). Maximal activities for EF and EF-H577N were 1208 s±1 and 4 s±1 (C) and those for EF and EF-H577N were 2726 s±1 and 6 s±1 (D), respectively. Means 6 SE are representative of at least two experiments. mutation at site 4 of CaM had a signi®cant effect on the switch C (Figure 1C). Based on sequence comparison, the EC50 value for CaM activation of CyaA-N with little adenylyl cyclase domain of CyaA-N (aa 1±393) does not change in Vmax value. Mutations at the other three sites contain sequences corresponding to the helical domain of caused little change (Figure 4D). EF, nor does ExoY, a related adenylyl cyclase from We puri®ed N- and C-terminal domains of CaM Pseudomonas aeruginosa (Yahr et al., 1998). However, an (N-CaM and C-CaM) and examined their abilities to EF mutant, EF-N (aa 291±640) that contains only the activate EF and CyaA-N (Figure 6A and B). We found that catalytic core and part of switch C had 10-fold reduction in either N-CaM or C-CaM alone could fully activate basal activity and 20-fold reduction in CaM activation, CyaA-N, although the EC50 values were increased highlighting the importance of the helical domain in the (Figure 6C and D). Similar to wild-type CaM, C-CaM activity of EF and its activation by CaM (Drum et al., can optimally activate CyaA-N in resting physiological 2000). This interpretation was compromised by the calcium concentrations (0.03±0.1 mM; Figure 6D). subsequent ®nding that the portion of switch C that was However, the activation of CyaA-N by N-CaM required deleted in EF-N (aa 641±655) is involved in CaM binding the elevated calcium concentrations (0.1±1 mM; (Drum et al., 2002). Thus, we made EF-DH (aa 291±655) Figure 6D). We also found that N-CaM could partially that had both the catalytic core and the entire switch C and activate EF (5% of wild-type CaM) and this activation was examined its ability to be activated by CaM (Figure 6A independent of the physiological concentrations of cal- and B). EF-DH was tagged with His6 at the C-terminal end cium (Figure 6B, E and F). We observed similar partial and was puri®ed using a Ni-NTA column to ensure that activation when CaM85/112 was used (Figure 4A). switch C was present in the protein. We found that, like C-CaM had signi®cantly reduced ability to activate EF EF-N, EF-DH had ~50-fold reduction in basal activity and and the addition of C-CaM did not further enhance the ~10-fold reduction in its activation by wild-type CaM. activation of EF by N-CaM (Figure 6B and E). The helical domain of EF makes substantial contact with the CA domain and switch C in the absence of CaM The role of the EF helical domain in EF activity and (combined contact surface = 3600 AÊ 2) and most of this CaM activation contact is disrupted upon CaM binding (Figure 2C and D; The adenylyl cyclase domain of EF consists of a catalytic Drum et al., 2002). Thus, we hypothesize that one of the core (CA and CB) and a helical domain that are linked by roles for the helical domain is to lock EF in an inactive

6727 Y.Shen et al.

Fig. 6. The activation of EF and CyaA-N by wild-type CaM, N- and C-terminal domain of CaM (N-CaM and C-CaM). (A) Coomassie Blue staining of puri®ed EF, EF mutants, CyaA-N, CaM and CaM mutants (2 mg each) on SDS±PAGE. (B) EF (1 nM), EF-DH (80 nM) and CyaA-N (0.7 nM) were assayed for adenylyl cyclase activity at 1.0 mM free Ca2+ and 10 mM CaM in the absence or the presence of full-length CaM, N-CaM, C-CaM or both N-CaM and C-CaM (N+C CaM). (C and E) Activation of CyaA-N and EF by wild-type and mutant CaM. Adenylyl cyclase toxin (0.7 nM CyaA-N or 1 nM EF) and 1.0 mM free Ca2+ were used in an adenylyl cyclase activity assay in the presence of wild-type CaM (®lled circles), C-CaM (open circles) and N-CaM (®lled triangles). (D and F) Calcium titration assay in CaM activation of CyaA-N and EF. Adenylyl cyclase activity of CyaA-N (0.7 nM) was measured with 10 mM wild-type CaM (®lled circles), C-CaM (open circles) and N-CaM (®lled triangles) and that of EF (1 nM) was examined with 1 mM wild-type CaM (®lled circles) and 10 mM N-CaM (®lled triangles). Maximal activity for wild-type CaM, N-CaM, and C-CaM is 1465, 230 and 1219 s±1, respectively. Means 6 SE are representative of at least two experiments.

state. The insertion of CaM between CA and the helical with CA and switch C but does not alter CaM binding domain leads to the extensive interaction between C-CaM should be more sensitive to CaM activation than wild-type with switch A and C. Such interaction triggers conforma- EF, since less energy is needed for CaM insertion. Only a tional changes in switch C which then stabilize the few residues, including E609, R613 and E616, ®t this catalytic loop (switch B) to achieve over 1000-fold criterion since most residues that make contact between catalytic activation (Figure 2C). This hypothesis predicts the CA and the helical domain are also involved in CaM that wild-type EF will be activated by C-CaM poorly since interaction (Figure 7A and B). We thus made an EF mutant C-CaM cannot insert itself between CA and the helical that had a single point mutation at E609, R613 or E616 to domain, while EF without the helical domain can interact alanine (EF-E609A, EF-R613A or EF-R616A) and a with C-CaM thus be better activated by C-CaM. mutant that had all three mutations (EF-triple). We found Consistent with this prediction, our result showed that that the adenylyl cyclase activity of C-CaM could activate EF-DH ~50-fold, but could only lysates that contained EF-E609A, EF-R613A, EF-E616A activate wild-type EF 3-fold (Figure 6B). and EF-triple mutants had 0.5-, 2-, 4- and 7-fold higher Our model also predicts that an EF mutant with a adenylyl cyclase activity, respectively, than one contain- mutation that disrupts the interaction of the helical domain ing wild-type EF (data not shown). We puri®ed EF-triple

6728 Activation of EF and CyaA by Ca2+±calmodulin

Fig. 7. Calcium±CaM activation of EF and EF-triple mutant. The locations of E609, R613 and E616 and their interacting residues, R718 and N633, are shown in the ribbon representation of EF structure in the absence (A) and presence (B) of CaM. CA±CB domain, switch C, helical domain and CaM are colored in green, purple, yellow and red, respectively. Ca2+ titration assay in the presence of 10 mM CaM (C) and CaM titration assay at 0.03, 1 and 10 mM calcium (D) were performed with 1 nM EF and 0.5 nM of EF-triple mutant. and found that EF-triple had 2- to 8-fold higher adenylyl requires physiological concentrations of calcium ions to cyclase activity than wild-type EF at 0.001±0.1 mM bind CaM at low CaM concentration (1 nM) and the calcium (Figures 6A and 7C). At relatively low calcium calcium-loaded CaM has higher af®nity for CyaA and EF concentrations, the EF-triple mutant had increased af®ni- than apo-CaM (Jurado et al., 1999). ties for CaM based on both SPR and kinetic analyses CyaA has 100-fold higher af®nity to CaM than EF does. (Figures 3D and 7D). This result is consistent with our Our study and the previous characterization of these two hypothesis that the interaction between the CA and the toxins reveal signi®cant differences in binding and acti- helical domain locks EF in the inactive state. vation of EF and CyaA by CaM. CyaA can be fully activated by CaM mutants with one defective C-terminal Ca2+-binding site or by the N- or C-terminal domain of Discussion CaM while EF cannot. A small region of CyaA In response to the elevation of intracellular Ca2+, CaM can (aa 225±267) contributes over 80% of binding free energy bind to and modulate the activity of numerous cellular (Bouhss et al., 1993) and the region corresponding to the proteins (Eldik and Watterson, 1998; Jurado et al., 1999). helical domain of EF is not required for its activation by Its ubiquitous expression and relative abundance in CaM. However, a large binding surface from four discrete Ê 2 eukaryotic cells make CaM ideally suited to be the regions of EF (~3000 A ) makes contact to CaM and the activator of anthrax and pertussis adenylyl cyclase toxins, helical domain is a major part of such interaction (Drum EF and CyaA, ensuring that these toxins are only active et al., 2002). inside the host cells. Proteins that interact with CaM can be Our biochemical and structural analysis in conjunction divided into two broad classes; those that have higher with a recent NMR study using EF in complex with 15N- af®nity for Ca2+-loaded CaM than for apo-CaM and those and 13C-labeled CaM reveals a new insight into the that prefer to bind apo-CaM. The ®rst class is exempli®ed dynamic interaction among EF, CaM and calcium ions by CaM kinase, calcineurin, cAMP phosphodiesterase and (Drum et al., 2002; T.Ulmer, S.Soelaiman, W.-J.Tang and type I adenylyl cyclase, while the second class is A.Bax, submitted). Our structural models of EF and exempli®ed by brush-border myosin I and neuromodulin EF±CaM reveal that the N-terminal domain of CaM (GAP-43). Although CyaA was previously categorized as interacts with the helical domain of EF, which is solvent a member of the second class based on its ability to bind exposed in the EF alone structure, while the C-terminal CaM in the presence of millimolar concentrations of domain of CaM interacts with several regions of EF that EGTA, our results using the combination of SPR, kinetic are mostly buried in the EF alone structure. This suggests and mutational analyses clearly show that CyaA, like EF, that CaM most likely makes the initial contact with EF via

6729 Y.Shen et al. its N-terminal domain. Consistent with this notion, our concentrations of its host cells. Consistent with this notion, mutational analysis shows that N-CaM, not C-CaM, can a recent study using CHO cells, neutrophils, macrophages partially activate EF and that this activation is independent and lymphocytes shows that the extracellular calcium will of calcium concentration. The NMR study further supports enter into the host cells upon the intoxication of edema this notion by showing that the N-terminal, but not toxin, and such calcium entry is required for the optimal C-terminal domain, of CaM makes contact with EF in the cAMP production by EF (Kumar et al., 2002). This study absence of calcium (T.Ulmer, S.Soelaiman, W.-J.Tang also shows that intoxication with edema toxin induces a and A.Bax, submitted). transient calcium spike in human lymphocytes, consistent Based on our structural models, we hypothesize that the with the notion that increases in intracellular cAMP can initial contact between N-CaM and the helical domain of elevate intracellular calcium concentration (Cooper et al., 2+ EF leads to the insertion of C-CaM between CA and the 1995). This cAMP-induced rise of Ca might then further helical domain. Such insertion triggers the conformational enhance the activity of EF in its host cells. changes of switch C, which stabilizes the catalytic loop, cAMP synthesis by adenylyl cyclase and degradation by leading to catalytic activation. Consistent with this notion, phosphodiesterase are tightly controlled in mammalian the mutations that weaken the interaction between CA and cells (Hanoune and Defer, 2001). Similar to the observa- the helical domain results in an EF mutant, EF-triple, that tion that the elevated intracellular Ca2+ can inhibit the has higher sensitivity to Ca2+±CaM. The interaction of enzymatic activity of adenylyl cyclases (type V and VI; C-CaM with the helical domain and CA of EF requires Cooper et al., 1995; Hu et al., 2002), we observe inhibition calcium binding to the C-terminal domain of CaM as seen of EF and CyaA by elevated calcium concentrations 2+ in the X-ray crystallographic structure of EF±CaM (Drum (IC50 = 0.3±1.0 mM). Although the role of Ca inhibition et al., 2002) and NMR analysis (T.Ulmer, S.Soelaiman, of EF and CyaA remains elusive, one might speculate that W.-J.Tang and A.Bax, submitted). Interestingly, the intracellular Ca2+ can control the rate of cAMP production N-terminal domain of CaM in the EF±CaM complex of adenylyl cyclase toxins. This control may reduce ATP can dynamically interact with calcium ions. The recent deprivation by adenylyl cyclase, which could lead to cell NMR analysis reveals that the addition of two more apoptosis and necrosis (McClintock et al., 2002). This moles of calcium allows calcium to bind the N-terminal could also set up the cAMP wave that is predicted to domain of CaM to promote further interaction between optimize cAMP-mediated signaling (Cooper et al., 1995). the N-terminal domain of CaM and EF (T.Ulmer, EF is secreted by B.anthracis, and spores of this S.Soelaiman, W.-J.Tang and A.Bax, submitted) even organism were used in the 2001 anthrax attacks in the US though no calcium ion is bound to the N-terminal domain (Inglesby et al., 2002). This incident has heightened of CaM in the X-ray crystallographic structure of the the need for anti-anthrax pharmaceuticals. Immobilizing EF±CaM complex. However, the calcium loading to the CaM±EF complex on a glass surface and determining the N-terminal domain of CaM does not affect the ability of structural basis for the ¯uorescence enhancement of CaM to activate EF based on the fact that neither the CaM 2¢d3¢ANT±ATP on binding to EF±CaM provides a viable mutant with a mutated calcium-binding site at the means by which to screen compounds that can block either N-terminal domain nor that with mutation to lock CaM activation of, or nucleotide binding to, EF. Adenylyl N-terminal domain in the calcium-free conformation has cyclase toxins are known to be secreted by two other reduced ability to activate EF. human pathogens, B.pertussis and P.aeruginosa, which Both EF and CyaA have about three orders of magni- cause whooping cough and 10±20% of hospital-acquired tude higher adenylyl cyclase activity than those of host , respectively (Yahr et al., 1998; Ladant and cells (Drum et al., 2002), thus the entry of EF and CyaA Ullmann, 1999). Biochemical studies have shown that should raise the cAMP concentration of host cells to the Yersinia pestis, the bacterium that causes plague also supra-physiological level. In this paper, we show that the secretes an adenylyl cyclase toxin and genome sequencing optimal activation of both EF and CyaA by CaM are identi®ed a gene in Y.pestis for such a toxin (Shevchenko highly active at the resting intracellular calcium concen- and Mishankin, 1987; Michankin et al., 1992; Parkhill tration when CaM is not limited (i.e. 10 mM). However, et al., 2001). Thus, adenylyl cyclase toxins are shared there is a major difference between these two toxins at among several potent human pathogens and the develop- relatively low free CaM concentrations (i.e. 0.1 mM). Our ment of therapeutic agents that block the activity of data show that the apparent af®nity of EF for CaM is low adenylyl cyclase toxins could have a broader usage against (Kd >10 mM) at the resting calcium concentration infections with pathogenic bacteria. (20±50 nM) so that EF may be minimally activated. EF becomes tightly associated with CaM (apparent Kd =5± 20 nM) and is fully active only when the intracellular Ca2+ Materials and methods is elevated to 1 mM. In contrast, CyaA has 100-fold higher af®nity to CaM than EF so that CyaA is still optimally Plasmid construction of EF mutants and CyaA-N mutants activated at the resting calcium concentration. Physio- The desired mutations were introduced using the Stratagene QuikChange Kit and the resulting mutations were con®rmed by DNA sequencing. To logical calcium concentrations only affect the activation of construct the plasmid for the expression of EF-triple, pProExH6-EF was CyaA-N by CaM when the free CaM is reduced to the used as a template with primers that introduced three point mutations from concentration that is unlikely to occur in the host cells (i.e. E609, R613 and E616 to alanine. The C-terminal His6-tagged EF-DH 1 nM CaM; Gerendasy, 1999; Jurado et al., 1999). (aa 291±658) was made using pProEx-EF-CH6 as a template, and a NotI site was introduced from residue 659 to 661 to remove the coding region This raises the possibility that EF activity can be of the helical domain (aa 659±800). The plasmid for the expression of 2+ regulated by the intracellular Ca concentration while CyaA-N(aa 1±393) was generated by introducing a termination codon at CyaA is fully active regardless of the intracellular calcium residue 394 using pEX-CyaA-1±412 as a template (Bejerano et al., 1999).

6730 Activation of EF and CyaA by Ca2+±calmodulin

Puri®cation of recombinant proteins Supplementary data For all the EF mutants, plasmids carrying the desired mutation were Supplementary data are available at The EMBO Journal Online. transformed into RNaseE-de®cient E.coli BL21 Star (DE3) cells harboring pUBS520. The mutant proteins were puri®ed using DEAE, SP-Sepharose and Ni-NTA columns as described (Drum et al., 2000). The Acknowledgements yield for EF mutants was ~30 mg/l E.coli culture. To express CyaA-N, pEX-CyaA-N was transformed into E.coli B834 cells containing We are grateful to P.Gardner (HHMI, University of Chicago) for help pUBS520. The resulting cells were grown in a modi®ed T7 medium with oligonucleotide synthesis and DNA sequencing, to Dr Emanuel Hanski at Hebrew University±Hadassah Medical School for the gene with 50 mg/ml ampicillin and 25 mg/ml kanamycin at 24°CtoA600 = 0.4, induced by adding isopropyl-1-thiogalactopyranoside to a ®nal concen- encoding CyaA, and to Drs Keith Brister, Gary Navrotski, Bill Desmarais tration of 100 mM and harvested overnight post-induction. E.coli cells and Robert Henning at APS BioCars 14-BMC for their help in data collection. This research was supported by National Institute of Health were lyzed in T20b5N500P0.1 buffer [20 mM Tris±HCl pH 8.0, 5 mM b-mercaptoethanol, 0.1 mM phenylmethylsulfonyl ¯uoride (PMSF), GM53459, GM62548 and American Heart Association Established 500 mM NaCl] by lysozyme (0.1 mg/ml) and sonication. After the lysate Investigator Award to W.-J.T., Defense Advanced Research Projects was spun at 200 000 g for 30 min, the supernatant of E.coli lysate was Agency N00173±01-G010 and National Science Foundation DMR- loaded directly onto a Ni2+-NTA column that was equilibrated with 9808595 to M.M., National Institute of Health AR41637 to Z.G. and 2+ Welch Foundation C-1119 to K.B. Use of the Advanced Photon Source T20b5N100P0.1. The Ni -NTA column was then washed with three was supported by the U.S. Department of Energy, Of®ce of Basic Energy column volumes of T20b5N100P0.1 and with T20b5N100P0.1 containing Sciences, under contract No. W-31-109-ENG-38. 20 mM imidazole (pH 7). The His6-tagged CyaA-N was then eluted with buffer T20b5N100P0.1 containing 150 mM imidazole. The peak fractions were loaded onto a Q-Sepharose column after 5-fold dilution with T20D1P0.1 (20 mM Tris±HCl pH 8.0, 1 mM dithiothreitol, 0.1 mM PMSF) References and the proteins were eluted by NaCl gradient. The puri®ed CyaA-N was then concentrated to >10 mg/ml and stored at ±80°C. The yield for Babu,Y.S., Sack,J.S., Greenhough,T.J., Bugg,C.E., Means,A.R. and CyaA-N was ~100 mg/l culture. Human and fruit ¯y full-length CaM, Cook,W.J. (1985) Three-dimensional structure of calmodulin. human CaM mutants (N-CaM and C-CaM), human cross-linked CaM and Nature, 315, 37±40. fruit ¯y CaM mutants (B1Q, B2Q, B3Q and B4Q) were expressed in Barbato,G., Ikura,M., Kay,L.E., Pastor,R.W. and Bax,A. (1992) 15 bacteria and puri®ed as described previously (Maune et al., 1992; Huber Backbone dynamics of calmodulin studied by N relaxation using et al., 1996). The molecular weight of N-CaM and C-CaM was con®rmed inverse detected two-dimensional NMR spectroscopy: the central by LC/MS. helix is ¯exible. Biochemistry, 31, 5269±5278. Bejerano,M., Nisan,I., Ludwig,A., Goebel,W. and Hanski,E. (1999) Characterization of the C-terminal domain essential for toxic activity Structure determination and ¯uorescence measurement of of adenylate cyclase toxin. Mol. Microbiol., 31, 381±392. EF±CaM±2¢d3¢ANT-ATP complex Bers,D., Patton,C. and Nuccitelli,R. (1994) A practical guide to the 2¢d3¢ANT-ATP was synthesized by reacting 2¢deoxy-ATP with isatoic preparation of Ca buffers. In Nuccitelli,R. (ed.), Methods Cell anhydride as described previously (Hiratsuka, 1983). To determine the Biology±A Practical Guide to the Study of Ca2+ in Living Cells. structure of EF±CaM±2¢d3¢ANT-ATP, crystals of EF3-CH6±CaM Vol. 40. Academic Press, San Diego, CA, pp. 3±29. complex were grown using vapor diffusion, soaked with 2 mM Bouhss,A., Krin,E., Munier,H., Gilles,A.M., Danchin,A., Glaser,P. and 2¢d3¢ANT-ATP during cryoprotection, and frozen in liquid nitrogen as Barzu,O. (1993) Cooperative phenomena in binding and activation of described (Drum et al., 2001). Data were collected at 100 K at APS adenylate cyclase by calmodulin. J. Biol. Chem., Biocars 14-BM-C and processed with the program Denzo and Scalepack 268, 1690±1694. (Otwinowski and Minor, 1997). The initial phase was obtained by Cooper,D.M., Mons,N. and Karpen,J.W. (1995) Adenylyl cyclases and difference Fourier method using program CNS and the model of EF±CaM Ê the interaction between calcium and cAMP signalling. Nature, 374, complex. The 3.6 A model was re®ned using the programs Turbo-Frodo, 421±424. O and CNS. The coordinates for EF±CaM±2¢d3¢ANT-ATP have been Deisseroth,K., Heist,E.K. and Tsien,R.W. (1998) Translocation of deposited with the Protein Data Bank (accession code 1LVC). Steady- calmodulin to the nucleus supports CREB phosphorylation in state ¯uorescence emission spectra were recorded on a Jobin Yvon-Spex hippocampal neurons. Nature, 392, 198±202. Fluoromax-2 photon counting spectro¯uorometer (Edison, NJ). The DeMaria,C.D., Soong,T.W., Alseikhan,B.A., Alvania,R.S. and Yue,D.T. indicated free calcium concentrations in the ¯uorescence assays as well as (2001) Calmodulin bifurcates the local Ca2+ signal that modulates the subsequent SPR and adenylyl cyclase assays were obtained by P/Q-type Ca2+ channels. Nature, 411, 484±489. buffering with EGTA based on the MAXC program website, http://www. Drum,C.L., Yan,S.Z., Sarac,R., Mabuchi,Y., Beckingham,K., Bohm,A., stanford.edu/~cpatton/maxc.html (Bers et al., 1994) and the conditions Grabarek,Z. and Tang,W.J. (2000) An extended conformation of for each experiment are listed in the tables in the supplementary data. calmodulin induces interactions between the structural domains of adenylyl cyclase from Bacillus anthracis to promote catalysis. J. Biol. SPR spectroscopy Chem., 275, 36334±36340. SPR measurements were performed with BIAcore 1000 instrument at Drum,C.L., Shen,Y., Rice,P.A., Bohm,A. and Tang,W.J. (2001) 25°C. Cut-CaM fusion protein was immobilized on a self-assembled Crystallization and preliminary X-ray study of the edema factor monolayer as described (Hodneland et al., 2002). Following incubation in exotoxin adenylyl cyclase domain from Bacillus anthracis in the buffer X (10 mM Tris±HCl pH 7.0, 150 mM KCl, 1.0 mM EGTA, 10 mM presence of its activator, calmodulin. Acta Crystallogr. D Biol. 2+ MgCl2 and variable CaCl2 concentrations to reach the desired free Ca Crystallogr., 57, 1881±1884. concentrations), buffer X containing adenylyl cyclase proteins was passed Drum,C.L., Yan,S.Z., Bard,J., Shen,Y., Lu,D., Soelaiman,S., over the immobilized calmodulin with a ¯ow rate of 1 ml/min for 5 min to Grabarek,Z., Bohm,A. and Tang,W.-J. (2002) Structural basis for allow association. The monolayer was then washed with buffer X for the activation of anthrax adenylyl cyclase exotoxin by calmodulin. >10 min to allow dissociation. Following each use, EGTA solution Nature, 415, 396±402. (100 mM) was washed over the surface to completely dissociate the Eldik,L.V. and Watterson,D.M. (1998) Calmodulin and Signal complex. Protein±protein interactions were measured at three different Transduction. Academic Press, San Diego, CA. adenylyl cyclase concentrations ranging from 10 nM to 8 mM, and the Finn,B.E., Evenas,J., Drakenberg,T., Waltho,J.P., Thulin,E. and data were ®t into the conformational change model provided by the Forsen,S. (1995) Calcium-induced structural changes and domain manufacturer of the instrument. autonomy in calmodulin. Nat. Struct. Biol., 2, 777±783. Gentile,F., Knipling,L.G., Sackett,D.L. and Wolff,J. (1990) Invasive Adenylyl cyclase assay adenylyl cyclase of Bordetella pertussis. Physical, catalytic and toxic Adenylyl cyclase activities were measured after 10 min at 30°C in the properties. J. Biol. Chem., 265, 10686±10692. presence of 20 mM HEPES pH 7.2, 5 mM ATP with a trace amount of Gerendasy,D. (1999) Homeostatic tuning of Ca2+ signal transduction by 32 [ P]ATP, 10 mM EGTA, 0.1 M KCl, 10 mM MgCl2 (unless indicated) members of the calpacitin protein family. J. Neurosci. Res., 58, and the indicated free calcium concentrations buffered by 10 mM EGTA. 107±119. ATP and cAMP were separated by Dowex and alumina columns as Hanoune,J. and Defer,N. (2001) Regulation and role of adenylyl cyclase described previously (Drum et al., 2000). isoforms. Ann. Rev. Pharm. Toxicol., 41, 145±174.

6731 Y.Shen et al.

Hiratsuka,T. (1983) New ribose-modi®ed ¯uorescent analogs of adenine Yahr,T.L., Vallis,A.J., Hancock,M.K., Barbieri,J.T. and Frank,D.W. and guanine nucleotides available as substrates for various enzymes. (1998) ExoY, an adenylate cyclase secreted by the Pseudomonas Biochim. Biophys. Acta, 742, 496±508. aeruginosa type III system. Proc. Natl Acad. Sci. USA, 95, Hodneland,C.D., Lee,Y.-S., Min,D.-H. and Mrksich,M. (2002) Selective 13899±13904. immobilization of protein to self-assembled monolayers presenting active site directed capture ligands. Proc. Natl Acad. Sci. USA, 99, Received August 15, 2002; revised October 24, 2002; 5048±5052. accepted October 29, 2002 Hoe¯ich,K.P. and Ikura,M. (2002) Calmodulin in action: diversity in target recognition and activation mechanisms. Cell, 108, 739±742. Hoover,D.L., Friedlander,A.M., Rogers,L.C., Yoon,I.K., Warren,R.L. and Cross,A.S. (1994) Anthrax edema toxin differentially regulates lipopolysaccharide-induced production of tumor necrosis factor a and interleukin-6 by increasing intracellular cyclic AMP. Infect. Immun., 62, 4432±4439. Hu,B., Nakata,H., Gu,C., de Beer,T. and Cooper,D.M. (2002) A critical interplay between Ca2+-inhibition and activation by Mg2+ of AC5 revealed by mutants and chimeric constructs. J. Biol. Chem., 277, 33139±33147. Huber,P.A., El-Mezgueldi,M., Grabarek,Z., Slatter,D.A., Levine,B.A. and Marston,S.B. (1996) Multiple-sited interaction of caldesmon with Ca2+±calmodulin. Biochem. J., 316, 413±420. Inglesby,T.V. et al. (2002) Anthrax as a biological weapon, 2002: updated recommendations for management. JAMA, 287, 2236±2252. Jurado,L.A., Chockalingam,P.S. and Jarrett,H.W. (1999) Apocal- modulin. Physiol. Rev., 79, 661±682. Khelef,N., Zychlinsky,A. and Guiso,N. (1993) Bordetella pertussis induces apoptosis in macrophages: role of adenylate cyclase- . Infect. Immun., 61, 4064±4071. Kumar,P., Ahuja,N. and Bhatnagar,R. (2002) Anthrax edema factor requires in¯ux of calcium for inducing cyclic AMP toxicity in target cells. Infect. Immun., 70, 4997±5007. Labruyure,E., Mock,M., Ladant,D., Michelson,S., Gilles,A.M., Laoide,B. and Barzu,O. (1990) Characterization of ATP and calmodulin-binding properties of a truncated form of Bacillus anthracis adenylate cyclase. Biochemistry, 29, 4922±4928. Ladant,D. and Ullmann,A. (1999) Bordetella pertussis adenylate cyclase: a toxin with multiple talents. Trends Microbiol., 7, 172±176. Leppla,S.H. (1984) Bacillus anthracis calmodulin-dependent adenylate cyclase: chemical and enzymatic properties and interactions with eucaryotic cells. Adv. Cyclic Nucleotide Protein Phosphorylation Res., 17, 189±198. Maune,J.F., Klee,C.B. and Beckingham,K. (1992) Ca2+ binding and conformational change in two series of point mutations to the individual Ca2+-binding sites of calmodulin. J. Biol. Chem., 267, 5286±5295. McClintock,D.S., Santore,M.T., Lee,V.Y., Brunelle,J., Budinger,G.R., Zong,W.X., Thompson,C.B., Hay,N. and Chandel,N.S. (2002) Bcl-2 family members and functional electron transport chain regulate oxygen deprivation-induced cell death. Mol. Cell. Biol., 22, 94±104. Meador,W.E. and Quiocho,F.A. (2002) Man bites dog. Nat. Struct. Biol., 9, 156±158. Michankin,B.N., Chevchenko,L.A. and Asseeva,L.E. (1992) Adenylate- cyclase. Un eventuel facteur de la pathogenicite de Yersinia pestis. Bull. Soc. Pathol. Exot., 85, 17±21. Mock,M. and Fouet,A. (2001) Anthrax. Annu. Rev. Microbiol., 55, 647±671. Otwinowski,Z. and Minor,W. (1997) Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol., 276, 307±326. Parkhill,J. et al. (2001) Genome sequence of Yersinia pestis, the causative agent of plague. Nature, 413, 523±527. Sarfati,R.S., Kansal,V.K., Munier,H., Glaser,P., Gilles,A.M., Labruyere,E., Mock,M., Danchin,A. and Barzu,O. (1990) Binding of 3¢-anthraniloyl-2¢-deoxy-ATP to calmodulin-activated adenylate cyclase from Bordetella pertussis and Bacillus anthracis. J. Biol. Chem., 265, 18902±18906. Schumacher,M.A., Rivard,A.F., Bachinger,H.P. and Adelman,J.P. (2001) Structure of the gating domain of a Ca2+-activated K+ channel complexed with Ca2+/calmodulin. Nature, 410, 1120±1124. Shevchenko,L.A. and Mishankin,B.N. (1987) Adenylyl cyclase of the causative agent of plague: its puri®cation and properties. Zh Mikrobiol. Epidemiol. Immunobiol., 7, 59±63. Tan,R.Y., Mabuchi,Y. and Grabarek,Z. (1996) Blocking the Ca2+- induced conformational transitions in calmodulin with disul®de bonds. J. Biol. Chem., 271, 7479±7483. Weingart,C.L. and Weiss,A.A. (2000) Bordetella pertussis virulence factors affect by human neutrophils. Infect. Immun., 68, 1735±1739.

6732